300-445 Question 212
Select 3An enterprise network operations center (NOC) is using a network assurance tool to monitor and maintain the health of their network. Recently, the network assurance tool generated multiple alerts regarding high CPU utilization on specific core switches. However, the NOC team discovered that many of these alerts were not actionable and led to false positives. To improve the signal-to-noise ratio, which actions should the network team take?
- A
Implement dynamic thresholds for CPU utilization alerts based on historical performance data.
- B
Disable all CPU utilization alerts to reduce noise and focus on other metrics.
- C
Configure the network assurance tool to correlate CPU utilization alerts with traffic patterns and application performance.
- D
Increase the polling interval for CPU utilization metrics to reduce the number of alerts.
- E
Use machine learning-based anomaly detection to identify outliers instead of relying solely on static thresholds.
Show answer and explanation
Correct answers: A, C, E
Explanation
Improving the signal-to-noise ratio in network assurance involves refining alerts to focus on actionable insights. Using dynamic thresholds, correlation with traffic and application data, and machine learning-based anomaly detection helps in reducing false positives and identifying real issues more effectively. Disabling alerts or increasing polling intervals is not a best practice as it may lead to missed critical events or delayed responses.
- A. Correct.
Dynamic thresholds adjust based on historical data, reducing false positives and improving the quality of actionable alerts.
- B. Incorrect.
Disabling all CPU utilization alerts is not a recommended practice as it could lead to missing critical issues, compromising network performance.
- C. Correct.
Correlating CPU utilization alerts with traffic patterns and application performance helps identify whether the high utilization is truly problematic or expected under certain conditions.
- D. Incorrect.
Increasing the polling interval might reduce the number of alerts but could lead to delayed detection of critical issues, which is not ideal for enterprise network assurance.
- E. Correct.
Machine learning-based anomaly detection provides a more intelligent way to detect unusual patterns and reduce reliance on static thresholds, improving alert accuracy.